4. Sensor systems and instrumentation

Injury assessment in frontal impact testing depends entirely on the quality of the measurement chain. A Hybrid III or THOR dummy is, at its core, an anthropomorphic mechanical structure populated with transducers that convert physical quantities—acceleration, force, moment, rotation—into electrical signals that the data acquisition system can record. Understanding what each sensor type measures, where it is mounted, how it is oriented, and how its output is conditioned and interpreted is essential before any injury criterion (HIC, Nij, chest deflection, femur force) can be trusted. This chapter examines the four principal transducer families used in both dummies: accelerometers, load cells, potentiometers, and angular rate sensors.

Accelerometers: The Primary Injury Signal

Piezoresistive accelerometers dominate dummy instrumentation because they can measure sustained (DC) acceleration, which is required for velocity and displacement integration. Each accelerometer is a small block containing a seismic mass on a silicon beam; strain-induced resistance changes are converted to a voltage proportional to acceleration. The critical practical detail is the polarity and axis convention: in the SAE J211 sign convention used throughout automotive crash testing, positive acceleration in the dummy's forward direction (anterior-to-posterior, +X) is defined as the direction that would push the occupant forward relative to the vehicle—i.e., a deceleration of the vehicle appears as a positive head acceleration. Every channel's polarity must be verified against this convention before the test, because a reversed head accelerometer will produce a plausible-looking but sign-inverted HIC curve.

Typical locations in the Hybrid III include the head CG (three orthogonal axes, often a triaxial package), the upper and lower thorax (chest accelerometers), and the pelvis. THOR adds a more extensive array: head CG triaxial, upper and lower sternum, and additional spine accelerometers. A representative head accelerometer might have a range of ±500 g and a frequency response flat to several kHz; the mounting surface must be rigid and the cable routed with a strain-relief loop so that dummy motion does not load the connector.

Load Cells: Forces and Moments at Interfaces

Load cells in dummies are strain-gauge bridges bonded to a machined element that deforms under load. They measure the force and moment transmitted across a structural interface. The most consequential examples are the upper neck load cell (six-axis: three forces, three moments), the lower neck transducer in THOR, the femur load cells (axial force for femur injury reference values), and the lumbar spine load cell. Each six-axis cell has a defined local coordinate system and a calibration matrix that maps the six bridge outputs to Fx, Fy, Fz, Mx, My, Mz. Cross-axis sensitivity and moment-arm offsets mean that the reported force at the cell's measurement origin is not identical to the force at the anatomical joint; the calibration certificate specifies the measurement point, and injury criteria must be applied at the correct reference location.

Practical checks include zero-load offset verification (the cell should read near zero with the dummy unloaded and in its reference posture) and a known-load tap test to confirm the channel responds in the expected direction. A common failure mode is a partially seated connector producing intermittent high-resistance contacts, which appears as spiky, non-physical force traces during the event.

Potentiometers: Measuring Deflection and Displacement

Potentiometers are resistive displacement transducers. In the Hybrid III chest, a rotary potentiometer (or a string potentiometer in some configurations) measures sternum-to-spine deflection, the input to the chest deflection criterion. In THOR, the chest deflection measurement is more elaborate, using multiple displacement sensors to capture deflection at several sternal locations. Potentiometers are also used for knee-slider displacement and, in some setups, for belt-pretensioner or steering-column travel.

The key installation concern is mechanical preload and range: the potentiometer must be set within its electrical travel at the dummy's reference posture, with enough stroke in both directions to avoid bottoming out during the impact. A deflection channel that clips at its mechanical stop will under-report peak deflection and produce a falsely benign injury number. Verify by manually deflecting the chest through its expected range before the test and confirming a smooth, monotonic voltage change with no flat segments.

Angular Rate Sensors: Rotation and Head Kinematics

Angular rate sensors (gyroscopes) measure rotational velocity about one or more axes. They are increasingly important because rotational head kinematics drive brain injury metrics (such as BrIC) and because angular rate data complements accelerometer data for reconstructing head motion. THOR commonly carries angular rate sensors in the head, and both dummies may use them in the head or upper neck depending on the test protocol. Modern units are MEMS-based and output a voltage or digital signal proportional to angular velocity in deg/s; integration yields angle, but drift and offset must be managed.

Because angular rate sensors are sensitive to offset and temperature, a pre-test bias check is mandatory: with the dummy stationary, each gyro channel should read within a few tenths of a degree per second of zero. A persistent offset will integrate into a large false angle over the event duration. Mounting orientation must also be documented—a gyro mounted with its sensing axis reversed will report rotation with the wrong sign, corrupting any derived angular displacement.

Signal Conditioning, Filtering, and Channel Verification

Raw transducer outputs are not directly usable. Load cells require bridge excitation (typically 2–10 V) and amplification; accelerometers and gyros require their specified supply and may need offset nulling. All channels pass through anti-alias filtering before digitization. The SAE J211 standard defines channel frequency classes (CFC 60, CFC 180, CFC 600, CFC 1000) that specify the filter corner frequency and roll-off; head acceleration is typically processed at CFC 1000, while chest deflection and femur force use lower classes. Applying the wrong CFC class changes the peak value and therefore the computed injury criterion—a subtle but consequential error.

A practical pre-test channel verification sequence for a frontal sled or crash test looks like this:

1. Static zero check: dummy at rest, upright. Confirm each accelerometer reads the expected gravity component, each load cell reads near zero, each potentiometer sits mid-range, and each gyro reads near zero rate.
2. Polarity tap test: apply a gentle, known-direction push to the head (forward) and confirm the head X accelerometer output goes positive per the J211 convention; tap the chest inward and confirm deflection increases.
3. Range and clipping check: manually move the chest through its expected deflection range and confirm the potentiometer does not reach its mechanical or electrical limit.
4. Noise floor check: record a few seconds of quiescent data and confirm channel noise is well below the expected signal amplitude; excessive noise suggests a grounding or shielding problem.
5. Channel-to-physical mapping audit: cross-check the DAQ channel list against the dummy build sheet so that, for example, the upper neck Fz channel is not mistakenly assigned to the lower neck.

Interpreting Sensor Data in Context

Once acquired, the signals must be read together. Head acceleration feeds HIC; neck load cell forces and moments feed Nij; chest potentiometer deflection feeds the chest deflection criterion; femur load cells feed femur force limits. Each criterion has a defined processing window, filter class, and reference point, and the sensor's measurement location must match the criterion's definition. For example, the upper neck load cell measures at its own origin, and the moment must be transferred to the occipital condyle before Nij is computed. Similarly, chest deflection measured at the sternum is not identical to the maximum deflection anywhere on the chest, which is why THOR's multi-point measurement provides richer information than the Hybrid III's single-channel potentiometer.

Angular rate data deserves separate attention: it is often used to reconstruct head rotation for brain injury metrics, and its accuracy depends on correct bias removal and integration. A small uncorrected offset—say 0.5 deg/s over a 100 ms event—integrates to only 0.05 degrees and is negligible, but over a longer event or with a larger offset the error grows. Always record a pre-event baseline and subtract it before integration.

The instrumentation philosophy differs meaningfully between the two dummies. The Hybrid III carries a well-established, standardized sensor set optimized for the classical injury criteria, with a single chest deflection channel and a six-axis upper neck load cell. THOR adds sensors—lower neck load cell, multi-point chest deflection, additional accelerometers, and angular rate sensors—that support more biofidelic injury metrics. The measurement chain, however, is the same in principle: transducer, conditioning, filtering, digitization, verification, and criterion computation. Mastering that chain, and the verification discipline that keeps it honest, is what separates trustworthy injury assessment from numbers that merely look plausible.